Skip to content


Repository files navigation



This package provides an implementation of the inference pipeline of AlphaFold v2. For simplicity, we refer to this model as AlphaFold throughout the rest of this document.

We also provide:

  1. An implementation of AlphaFold-Multimer. This represents a work in progress and AlphaFold-Multimer isn't expected to be as stable as our monomer AlphaFold system. Read the guide for how to upgrade and update code.
  2. The technical note containing the models and inference procedure for an updated AlphaFold v2.3.0.
  3. A CASP15 baseline set of predictions along with documentation of any manual interventions performed.

Any publication that discloses findings arising from using this source code or the model parameters should cite the AlphaFold paper and, if applicable, the AlphaFold-Multimer paper.

Please also refer to the Supplementary Information for a detailed description of the method.

You can use a slightly simplified version of AlphaFold with this Colab notebook or community-supported versions (see below).

If you have any questions, please contact the AlphaFold team at

CASP14 predictions

Installation and running your first prediction

You will need a machine running Linux, AlphaFold does not support other operating systems. Full installation requires up to 3 TB of disk space to keep genetic databases (SSD storage is recommended) and a modern NVIDIA GPU (GPUs with more memory can predict larger protein structures).

Please follow these steps:

  1. Install Docker.

  2. Clone this repository and cd into it.

    git clone
    cd ./alphafold
  3. Download genetic databases and model parameters:

    • Install aria2c. On most Linux distributions it is available via the package manager as the aria2 package (on Debian-based distributions this can be installed by running sudo apt install aria2).

    • Please use the script scripts/ to download and set up full databases. This may take substantial time (download size is 556 GB), so we recommend running this script in the background:

    scripts/ <DOWNLOAD_DIR> > download.log 2> download_all.log &
    • Note: The download directory <DOWNLOAD_DIR> should not be a subdirectory in the AlphaFold repository directory. If it is, the Docker build will be slow as the large databases will be copied into the docker build context.

    • It is possible to run AlphaFold with reduced databases; please refer to the complete documentation.

  4. Check that AlphaFold will be able to use a GPU by running:

    docker run --rm --gpus all nvidia/cuda:11.0-base nvidia-smi

    The output of this command should show a list of your GPUs. If it doesn't, check if you followed all steps correctly when setting up the NVIDIA Container Toolkit or take a look at the following NVIDIA Docker issue.

    If you wish to run AlphaFold using Singularity (a common containerization platform on HPC systems) we recommend using some of the third party Singularity setups as linked in #10 or #24.

  5. Build the Docker image:

    docker build -f docker/Dockerfile -t alphafold .

    If you encounter the following error:

    W: GPG error: InRelease: The following signatures couldn't be verified because the public key is not available: NO_PUBKEY A4B469963BF863CC
    E: The repository ' InRelease' is not signed.

    use the workaround described in #463 (comment).

  6. Install the dependencies. Note: You may optionally wish to create a Python Virtual Environment to prevent conflicts with your system's Python environment.

    pip3 install -r docker/requirements.txt
  7. Make sure that the output directory exists (the default is /tmp/alphafold) and that you have sufficient permissions to write into it.

  8. Run pointing to a FASTA file containing the protein sequence(s) for which you wish to predict the structure (--fasta_paths parameter). AlphaFold will search for the available templates before the date specified by the --max_template_date parameter; this could be used to avoid certain templates during modeling. --data_dir is the directory with downloaded genetic databases and --output_dir is the absolute path to the output directory.

    python3 docker/ \
      --fasta_paths=your_protein.fasta \
      --max_template_date=2022-01-01 \
      --data_dir=$DOWNLOAD_DIR \
  9. Once the run is over, the output directory shall contain predicted structures of the target protein. Please check the documentation below for additional options and troubleshooting tips.

Genetic databases

This step requires aria2c to be installed on your machine.

AlphaFold needs multiple genetic (sequence) databases to run:

We provide a script scripts/ that can be used to download and set up all of these databases:

  • Recommended default:

    scripts/ <DOWNLOAD_DIR>

    will download the full databases.

  • With reduced_dbs parameter:

    scripts/ <DOWNLOAD_DIR> reduced_dbs

    will download a reduced version of the databases to be used with the reduced_dbs database preset. This shall be used with the corresponding AlphaFold parameter --db_preset=reduced_dbs later during the AlphaFold run (please see AlphaFold parameters section).

📒 Note: The download directory <DOWNLOAD_DIR> should not be a subdirectory in the AlphaFold repository directory. If it is, the Docker build will be slow as the large databases will be copied during the image creation.

We don't provide exactly the database versions used in CASP14 – see the note on reproducibility. Some of the databases are mirrored for speed, see mirrored databases.

📒 Note: The total download size for the full databases is around 556 GB and the total size when unzipped is 2.62 TB. Please make sure you have a large enough hard drive space, bandwidth and time to download. We recommend using an SSD for better genetic search performance.

📒 Note: If the download directory and datasets don't have full read and write permissions, it can cause errors with the MSA tools, with opaque (external) error messages. Please ensure the required permissions are applied, e.g. with the sudo chmod 755 --recursive "$DOWNLOAD_DIR" command.

The script will also download the model parameter files. Once the script has finished, you should have the following directory structure:

$DOWNLOAD_DIR/                             # Total: ~ 2.62 TB (download: 556 GB)
    bfd/                                   # ~ 1.8 TB (download: 271.6 GB)
        # 6 files.
    mgnify/                                # ~ 120 GB (download: 67 GB)
    params/                                # ~ 5.3 GB (download: 5.3 GB)
        # 5 CASP14 models,
        # 5 pTM models,
        # 5 AlphaFold-Multimer models,
        # LICENSE,
        # = 16 files.
    pdb70/                                 # ~ 56 GB (download: 19.5 GB)
        # 9 files.
    pdb_mmcif/                             # ~ 238 GB (download: 43 GB)
            # About 199,000 .cif files.
    pdb_seqres/                            # ~ 0.2 GB (download: 0.2 GB)
    small_bfd/                             # ~ 17 GB (download: 9.6 GB)
    uniref30/                              # ~ 206 GB (download: 52.5 GB)
        # 7 files.
    uniprot/                               # ~ 105 GB (download: 53 GB)
    uniref90/                              # ~ 67 GB (download: 34 GB)

bfd/ is only downloaded if you download the full databases, and small_bfd/ is only downloaded if you download the reduced databases.

Model parameters

While the AlphaFold code is licensed under the Apache 2.0 License, the AlphaFold parameters and CASP15 prediction data are made available under the terms of the CC BY 4.0 license. Please see the Disclaimer below for more detail.

The AlphaFold parameters are available from, and are downloaded as part of the scripts/ script. This script will download parameters for:

  • 5 models which were used during CASP14, and were extensively validated for structure prediction quality (see Jumper et al. 2021, Suppl. Methods 1.12 for details).
  • 5 pTM models, which were fine-tuned to produce pTM (predicted TM-score) and (PAE) predicted aligned error values alongside their structure predictions (see Jumper et al. 2021, Suppl. Methods 1.9.7 for details).
  • 5 AlphaFold-Multimer models that produce pTM and PAE values alongside their structure predictions.

Updating existing installation

If you have a previous version you can either reinstall fully from scratch (remove everything and run the setup from scratch) or you can do an incremental update that will be significantly faster but will require a bit more work. Make sure you follow these steps in the exact order they are listed below:

  1. Update the code.
    • Go to the directory with the cloned AlphaFold repository and run git fetch origin main to get all code updates.
  2. Update the UniProt, UniRef, MGnify and PDB seqres databases.
    • Remove <DOWNLOAD_DIR>/uniprot.
    • Run scripts/ <DOWNLOAD_DIR>.
    • Remove <DOWNLOAD_DIR>/uniclust30.
    • Run scripts/ <DOWNLOAD_DIR>.
    • Remove <DOWNLOAD_DIR>/uniref90.
    • Run scripts/ <DOWNLOAD_DIR>.
    • Remove <DOWNLOAD_DIR>/mgnify.
    • Run scripts/ <DOWNLOAD_DIR>.
    • Remove <DOWNLOAD_DIR>/pdb_mmcif. It is needed to have PDB SeqRes and PDB from exactly the same date. Failure to do this step will result in potential errors when searching for templates when running AlphaFold-Multimer.
    • Run scripts/ <DOWNLOAD_DIR>.
    • Run scripts/ <DOWNLOAD_DIR>.
  3. Update the model parameters.
    • Remove the old model parameters in <DOWNLOAD_DIR>/params.
    • Download new model parameters using scripts/ <DOWNLOAD_DIR>.
  4. Follow Running AlphaFold.

Using deprecated model weights

To use the deprecated v2.2.0 AlphaFold-Multimer model weights:

  1. Change SOURCE_URL in scripts/ to, and download the old parameters.
  2. Change the _v3 to _v2 in the multimer MODEL_PRESETS in

To use the deprecated v2.1.0 AlphaFold-Multimer model weights:

  1. Change SOURCE_URL in scripts/ to, and download the old parameters.
  2. Remove the _v3 in the multimer MODEL_PRESETS in

Running AlphaFold

The simplest way to run AlphaFold is using the provided Docker script. This was tested on Google Cloud with a machine using the nvidia-gpu-cloud-image with 12 vCPUs, 85 GB of RAM, a 100 GB boot disk, the databases on an additional 3 TB disk, and an A100 GPU. For your first run, please follow the instructions from Installation and running your first prediction section.

  1. By default, Alphafold will attempt to use all visible GPU devices. To use a subset, specify a comma-separated list of GPU UUID(s) or index(es) using the --gpu_devices flag. See GPU enumeration for more details.

  2. You can control which AlphaFold model to run by adding the --model_preset= flag. We provide the following models:

    • monomer: This is the original model used at CASP14 with no ensembling.

    • monomer_casp14: This is the original model used at CASP14 with num_ensemble=8, matching our CASP14 configuration. This is largely provided for reproducibility as it is 8x more computationally expensive for limited accuracy gain (+0.1 average GDT gain on CASP14 domains).

    • monomer_ptm: This is the original CASP14 model fine tuned with the pTM head, providing a pairwise confidence measure. It is slightly less accurate than the normal monomer model.

    • multimer: This is the AlphaFold-Multimer model. To use this model, provide a multi-sequence FASTA file. In addition, the UniProt database should have been downloaded.

  3. You can control MSA speed/quality tradeoff by adding --db_preset=reduced_dbs or --db_preset=full_dbs to the run command. We provide the following presets:

    • reduced_dbs: This preset is optimized for speed and lower hardware requirements. It runs with a reduced version of the BFD database. It requires 8 CPU cores (vCPUs), 8 GB of RAM, and 600 GB of disk space.

    • full_dbs: This runs with all genetic databases used at CASP14.

    Running the command above with the monomer model preset and the reduced_dbs data preset would look like this:

    python3 docker/ \
      --fasta_paths=T1050.fasta \
      --max_template_date=2020-05-14 \
      --model_preset=monomer \
      --db_preset=reduced_dbs \
      --data_dir=$DOWNLOAD_DIR \
  4. After generating the predicted model, AlphaFold runs a relaxation step to improve local geometry. By default, only the best model (by pLDDT) is relaxed (--models_to_relax=best), but also all of the models (--models_to_relax=all) or none of the models (--models_to_relax=none) can be relaxed.

  5. The relaxation step can be run on GPU (faster, but could be less stable) or CPU (slow, but stable). This can be controlled with --enable_gpu_relax=true (default) or --enable_gpu_relax=false.

  6. AlphaFold can re-use MSAs (multiple sequence alignments) for the same sequence via --use_precomputed_msas=true option; this can be useful for trying different AlphaFold parameters. This option assumes that the directory structure generated by the first AlphaFold run in the output directory exists and that the protein sequence is the same.

Running AlphaFold-Multimer

All steps are the same as when running the monomer system, but you will have to

  • provide an input fasta with multiple sequences,
  • set --model_preset=multimer,

An example that folds a protein complex multimer.fasta:

python3 docker/ \
  --fasta_paths=multimer.fasta \
  --max_template_date=2020-05-14 \
  --model_preset=multimer \
  --data_dir=$DOWNLOAD_DIR \

By default the multimer system will run 5 seeds per model (25 total predictions) for a small drop in accuracy you may wish to run a single seed per model. This can be done via the --num_multimer_predictions_per_model flag, e.g. set it to --num_multimer_predictions_per_model=1 to run a single seed per model.

AlphaFold prediction speed

The table below reports prediction runtimes for proteins of various lengths. We only measure unrelaxed structure prediction with three recycles while excluding runtimes from MSA and template search. When running docker/ with --benchmark=true, this runtime is stored in timings.json. All runtimes are from a single A100 NVIDIA GPU. Prediction speed on A100 for smaller structures can be improved by increasing global_config.subbatch_size in alphafold/model/

No. residues Prediction time (s)
100 4.9
200 7.7
300 13
400 18
500 29
600 36
700 53
800 60
900 91
1,000 96
1,100 140
1,500 280
2,000 450
2,500 969
3,000 1,240
3,500 2,465
4,000 5,660
4,500 12,475
5,000 18,824


Below are examples on how to use AlphaFold in different scenarios.

Folding a monomer

Say we have a monomer with the sequence <SEQUENCE>. The input fasta should be:


Then run the following command:

python3 docker/ \
  --fasta_paths=monomer.fasta \
  --max_template_date=2021-11-01 \
  --model_preset=monomer \
  --data_dir=$DOWNLOAD_DIR \

Folding a homomer

Say we have a homomer with 3 copies of the same sequence <SEQUENCE>. The input fasta should be:


Then run the following command:

python3 docker/ \
  --fasta_paths=homomer.fasta \
  --max_template_date=2021-11-01 \
  --model_preset=multimer \
  --data_dir=$DOWNLOAD_DIR \

Folding a heteromer

Say we have an A2B3 heteromer, i.e. with 2 copies of <SEQUENCE A> and 3 copies of <SEQUENCE B>. The input fasta should be:


Then run the following command:

python3 docker/ \
  --fasta_paths=heteromer.fasta \
  --max_template_date=2021-11-01 \
  --model_preset=multimer \
  --data_dir=$DOWNLOAD_DIR \

Folding multiple monomers one after another

Say we have a two monomers, monomer1.fasta and monomer2.fasta.

We can fold both sequentially by using the following command:

python3 docker/ \
  --fasta_paths=monomer1.fasta,monomer2.fasta \
  --max_template_date=2021-11-01 \
  --model_preset=monomer \
  --data_dir=$DOWNLOAD_DIR \

Folding multiple multimers one after another

Say we have a two multimers, multimer1.fasta and multimer2.fasta.

We can fold both sequentially by using the following command:

python3 docker/ \
  --fasta_paths=multimer1.fasta,multimer2.fasta \
  --max_template_date=2021-11-01 \
  --model_preset=multimer \
  --data_dir=$DOWNLOAD_DIR \

AlphaFold output

The outputs will be saved in a subdirectory of the directory provided via the --output_dir flag of (defaults to /tmp/alphafold/). The outputs include the computed MSAs, unrelaxed structures, relaxed structures, ranked structures, raw model outputs, prediction metadata, and section timings. The --output_dir directory will have the following structure:


The contents of each output file are as follows:

  • features.pkl – A pickle file containing the input feature NumPy arrays used by the models to produce the structures.

  • unrelaxed_model_*.pdb – A PDB format text file containing the predicted structure, exactly as outputted by the model.

  • relaxed_model_*.pdb – A PDB format text file containing the predicted structure, after performing an Amber relaxation procedure on the unrelaxed structure prediction (see Jumper et al. 2021, Suppl. Methods 1.8.6 for details).

  • ranked_*.pdb – A PDB format text file containing the predicted structures, after reordering by model confidence. Here ranked_i.pdb should contain the prediction with the (i + 1)-th highest confidence (so that ranked_0.pdb has the highest confidence). To rank model confidence, we use predicted LDDT (pLDDT) scores (see Jumper et al. 2021, Suppl. Methods 1.9.6 for details). If --models_to_relax=all then all ranked structures are relaxed. If --models_to_relax=best then only ranked_0.pdb is relaxed (the rest are unrelaxed). If --models_to_relax=none, then the ranked structures are all unrelaxed.

  • ranking_debug.json – A JSON format text file containing the pLDDT values used to perform the model ranking, and a mapping back to the original model names.

  • relax_metrics.json – A JSON format text file containing relax metrics, for instance remaining violations.

  • timings.json – A JSON format text file containing the times taken to run each section of the AlphaFold pipeline.

  • msas/ - A directory containing the files describing the various genetic tool hits that were used to construct the input MSA.

  • result_model_*.pkl – A pickle file containing a nested dictionary of the various NumPy arrays directly produced by the model. In addition to the output of the structure module, this includes auxiliary outputs such as:

    • Distograms (distogram/logits contains a NumPy array of shape [N_res, N_res, N_bins] and distogram/bin_edges contains the definition of the bins).
    • Per-residue pLDDT scores (plddt contains a NumPy array of shape [N_res] with the range of possible values from 0 to 100, where 100 means most confident). This can serve to identify sequence regions predicted with high confidence or as an overall per-target confidence score when averaged across residues.
    • Present only if using pTM models: predicted TM-score (ptm field contains a scalar). As a predictor of a global superposition metric, this score is designed to also assess whether the model is confident in the overall domain packing.
    • Present only if using pTM models: predicted pairwise aligned errors (predicted_aligned_error contains a NumPy array of shape [N_res, N_res] with the range of possible values from 0 to max_predicted_aligned_error, where 0 means most confident). This can serve for a visualisation of domain packing confidence within the structure.

The pLDDT confidence measure is stored in the B-factor field of the output PDB files (although unlike a B-factor, higher pLDDT is better, so care must be taken when using for tasks such as molecular replacement).

This code has been tested to match mean top-1 accuracy on a CASP14 test set with pLDDT ranking over 5 model predictions (some CASP targets were run with earlier versions of AlphaFold and some had manual interventions; see our forthcoming publication for details). Some targets such as T1064 may also have high individual run variance over random seeds.

Inferencing many proteins

The provided inference script is optimized for predicting the structure of a single protein, and it will compile the neural network to be specialized to exactly the size of the sequence, MSA, and templates. For large proteins, the compile time is a negligible fraction of the runtime, but it may become more significant for small proteins or if the multi-sequence alignments are already precomputed. In the bulk inference case, it may make sense to use our make_fixed_size function to pad the inputs to a uniform size, thereby reducing the number of compilations required.

We do not provide a bulk inference script, but it should be straightforward to develop on top of the RunModel.predict method with a parallel system for precomputing multi-sequence alignments. Alternatively, this script can be run repeatedly with only moderate overhead.

Note on CASP14 reproducibility

AlphaFold's output for a small number of proteins has high inter-run variance, and may be affected by changes in the input data. The CASP14 target T1064 is a notable example; the large number of SARS-CoV-2-related sequences recently deposited changes its MSA significantly. This variability is somewhat mitigated by the model selection process; running 5 models and taking the most confident.

To reproduce the results of our CASP14 system as closely as possible you must use the same database versions we used in CASP. These may not match the default versions downloaded by our scripts.

For genetics:

For templates:

An alternative for templates is to use the latest PDB and PDB70, but pass the flag --max_template_date=2020-05-14, which restricts templates only to structures that were available at the start of CASP14.

Citing this work

If you use the code or data in this package, please cite:

  author  = {Jumper, John and Evans, Richard and Pritzel, Alexander and Green, Tim and Figurnov, Michael and Ronneberger, Olaf and Tunyasuvunakool, Kathryn and Bates, Russ and {\v{Z}}{\'\i}dek, Augustin and Potapenko, Anna and Bridgland, Alex and Meyer, Clemens and Kohl, Simon A A and Ballard, Andrew J and Cowie, Andrew and Romera-Paredes, Bernardino and Nikolov, Stanislav and Jain, Rishub and Adler, Jonas and Back, Trevor and Petersen, Stig and Reiman, David and Clancy, Ellen and Zielinski, Michal and Steinegger, Martin and Pacholska, Michalina and Berghammer, Tamas and Bodenstein, Sebastian and Silver, David and Vinyals, Oriol and Senior, Andrew W and Kavukcuoglu, Koray and Kohli, Pushmeet and Hassabis, Demis},
  journal = {Nature},
  title   = {Highly accurate protein structure prediction with {AlphaFold}},
  year    = {2021},
  volume  = {596},
  number  = {7873},
  pages   = {583--589},
  doi     = {10.1038/s41586-021-03819-2}

In addition, if you use the AlphaFold-Multimer mode, please cite:

@article {AlphaFold-Multimer2021,
  author       = {Evans, Richard and O{\textquoteright}Neill, Michael and Pritzel, Alexander and Antropova, Natasha and Senior, Andrew and Green, Tim and {\v{Z}}{\'\i}dek, Augustin and Bates, Russ and Blackwell, Sam and Yim, Jason and Ronneberger, Olaf and Bodenstein, Sebastian and Zielinski, Michal and Bridgland, Alex and Potapenko, Anna and Cowie, Andrew and Tunyasuvunakool, Kathryn and Jain, Rishub and Clancy, Ellen and Kohli, Pushmeet and Jumper, John and Hassabis, Demis},
  journal      = {bioRxiv},
  title        = {Protein complex prediction with AlphaFold-Multimer},
  year         = {2021},
  elocation-id = {2021.10.04.463034},
  doi          = {10.1101/2021.10.04.463034},
  URL          = {},
  eprint       = {},

Community contributions

Colab notebooks provided by the community (please note that these notebooks may vary from our full AlphaFold system and we did not validate their accuracy):


AlphaFold communicates with and/or references the following separate libraries and packages:

We thank all their contributors and maintainers!

Get in Touch

If you have any questions not covered in this overview, please contact the AlphaFold team at

We would love to hear your feedback and understand how AlphaFold has been useful in your research. Share your stories with us at

License and Disclaimer

This is not an officially supported Google product.

Copyright 2022 DeepMind Technologies Limited.

AlphaFold Code License

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.

Model Parameters License

The AlphaFold parameters are made available under the terms of the Creative Commons Attribution 4.0 International (CC BY 4.0) license. You can find details at:

Third-party software

Use of the third-party software, libraries or code referred to in the Acknowledgements section above may be governed by separate terms and conditions or license provisions. Your use of the third-party software, libraries or code is subject to any such terms and you should check that you can comply with any applicable restrictions or terms and conditions before use.

Mirrored Databases

The following databases have been mirrored by DeepMind, and are available with reference to the following: